Ship target plot-track association method and system based on graph representation learning

Abstract

The invention discloses a graph representation learning-based ship target plot-track association method and system, and relates to the technical field of radar target tracking and data association, and the method comprises the steps: obtaining an R-D spectrum based on an obtained echo signal, dividing a target foreground, and constructing an undirected weighted graph structure; converting the graph structure into low-dimensional potential embedding by using an encoder, and reconstructing an adjacent matrix by using a decoder; after training is completed, feature extraction is conducted on the graph structure through an encoder, and graph feature vectors are obtained; and at the trace point-track association gate, respectively calculating the similarity between the graph feature vectors of the candidate measurement trace points and the graph feature vectors of the three historical frames of the current track, and selecting the candidate measurement trace point with the highest similarity as the trace point associated with the current track. According to the method, accurate correlation between the trace point and the track is realized, and the problem of correlation errors caused by low signal-to-noise ratio and single target feature is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of radar target tracking and data association, and in particular to a method and system for ship target point-track association based on graph representation learning. Background Art

[0002] With the continuous improvement of the demand for maritime surveillance, compact ground-wave radars have been widely used in the field of maritime ship target detection due to their advantages such as small size, low cost, and flexible deployment. However, due to system limitations (such as low transmission power and low azimuth resolution) and the influence of interference such as sea clutter and ground clutter, radar echoes often have problems such as low signal-to-noise ratio and high measurement uncertainty. Traditional maritime ship point-track association methods mainly rely on target kinematic parameters, and it is often difficult to distinguish adjacent targets and clutter in complex environments, easily causing point-track association ambiguity problems, resulting in frequent incorrect associations or track breaks.

[0003] With the continuous development of deep learning technology, in recent years, in some studies, attempts have been made to introduce deep learning technology into the field of data association. However, compact ground-wave radars are severely affected by complex working environments and variable target characteristics, and still face the problem of weak target features in the point-track association process. Existing target detection frameworks often only focus on extracting the kinematic parameters of targets, ignoring the rich spectral features in radar echoes. In fact, the range-Doppler (R-D) spectrum of radar can not only reflect the geometric position and velocity information of targets, but also contain rich energy distribution and spatial correlation data, which is of great value for distinguishing different targets and targets from clutter. However, due to the complexity of electromagnetic interactions, traditional signal processing methods are difficult to effectively extract and utilize these spectral features. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the present invention provides a method and system for ship target point-track association based on graph representation learning, which uses graph representation learning to extract graph structure features reflecting target scattering characteristics from the R-D spectrum, constructs recognizable target feature vectors, thereby realizing accurate association of points and tracks, and overcoming the association error problems caused by low signal-to-noise ratio and single target features.

[0005] To achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a method for ship target point-track association based on graph representation learning, including: Obtain an echo signal containing ship target data and maritime environment data, and preprocess the echo signal to obtain an R-D spectrum; Based on the R-D spectrum, the target foreground is segmented, and the cells in the target foreground are used as nodes to construct an undirected weighted graph structure including a node feature matrix and an adjacency matrix; The encoder is used to convert the graph structure into a low-dimensional latent embedding, and then the decoder is used to reconstruct the adjacency matrix according to the latent embedding; the encoder and the decoder are trained, and after the training is completed, the encoder is used to extract features from the graph structure to obtain a graph feature vector; The graph feature vectors of the candidate measurement point tracks and the graph feature vectors of the last three frames of the current track history are obtained respectively. At the point track-track association gate, the similarities between the graph feature vectors of the candidate measurement point tracks and the graph feature vectors of the last three frames of the current track history are calculated respectively, and the candidate measurement point track with the highest similarity is selected as the point track associated with the current track.

[0006] As a further technical solution, the preprocessing includes: first, the echo signal is converted into a digital signal through A / D conversion, and then through pulse compression, beamforming, and Doppler processing, it is converted into an R-D spectrum; each resolution cell of the R-D spectrum is represented by a triple composed of distance, Doppler velocity, and echo amplitude.

[0007] As a further technical solution, the method for segmenting the target foreground is: in the R-D spectrum, a window is selected with the cell where the target peak is located as the center, and the Otsu method is used to determine the optimal amplitude segmentation threshold between the foreground and the background within the window; a target foreground segmentation mask is generated based on the optimal amplitude segmentation threshold, and the cells with an amplitude greater than or equal to the optimal amplitude segmentation threshold are segmented into the target foreground.

[0008] As a further technical solution, the method for constructing an undirected weighted graph structure is: first, the node corresponding to the target peak cell in the target foreground is set as the central node and numbered "1", and other nodes are numbered in the order of their relative positions to the central node; then the central node is connected to other nodes, and finally an undirected weighted graph structure is constructed.

[0009] As a further technical solution, in the training stage, the graph structure is input into the encoder, and the encoder uses a graph convolutional network to convert the graph structure into a low-dimensional latent embedding, and its encoding process can be expressed as: ; where is the low-dimensional latent embedding, is the graph convolutional network, is the node feature matrix, is the adjacency matrix; The formula for the decoder to reconstruct the adjacency matrix according to the latent embedding is: ; where is the reconstructed adjacency matrix, is the activation function.

[0010] As a further technical solution, the encoder adopts a two-layer GCN architecture; after training is completed, first, the first-layer GCN converts the input graph structure into a hidden embedding through a message passing operation, and then uses the Leaky ReLU activation function to introduce non-linearity; based on the hidden embedding generated by the first layer, the second-layer GCN executes the message passing mechanism again to generate node embeddings; finally, through a global average pooling operation, the feature vectors of all nodes are aggregated to obtain a graph feature vector.

[0011] As a further technical solution, calculate the similarity between the graph feature vector of the candidate measurement point track and the graph feature vectors of the three frames of the current track history respectively. The specific method is: calculate the cosine similarity between the graph feature vector of the candidate measurement point track and the graph feature vectors of the three frames of the current track history respectively, and then obtain the comprehensive similarity through weighted fusion. The calculation formula is: ; where , , are weight coefficients, is the graph feature vector of the candidate measurement point track, , , are the graph feature vectors of the three frames of the current track history.

[0012] In the second aspect, the present invention provides a vessel target point track - track association system based on graph representation learning, including the following modules: Data acquisition and preprocessing module, configured to: acquire an echo signal containing vessel target data and marine environment data, and preprocess the echo signal to obtain an R - D spectrum; Graph structure construction module, configured to: based on the R - D spectrum, divide the target foreground, and use the cells in the target foreground as nodes to construct an undirected weighted graph structure including a node feature matrix and an adjacency matrix; Feature extraction module, configured to: use the encoder to convert the graph structure into a low - dimensional latent embedding, and then use the decoder to reconstruct the adjacency matrix according to the latent embedding; train the encoder and the decoder, and after training is completed, use the encoder to extract features from the graph structure to obtain a graph feature vector; Point track - track association module, configured to: respectively obtain the graph feature vectors of the candidate measurement point track and the graph feature vectors of the three frames of the current track history, and at the point track - track association gate, calculate the similarity between the graph feature vector of the candidate measurement point track and the graph feature vectors of the three frames of the current track history respectively, and select the candidate measurement point track with the highest similarity as the point track associated with the current track.

[0013] One or more technical solutions of the present invention have the following beneficial effects: 1. By acquiring echo signals, delving into the spectral information in the echo signals, and extracting the energy distribution and spatial structure of the targets in the R-D spectrum as an important basis for correlation judgment, the present invention provides richer and more discriminative information support for differentiating different traces, significantly improving the accuracy and reliability of trace-to-track correlation.

[0014] 2. By extracting graph feature vectors, the present invention converts spectral features into effective information available for correlation. Specifically, by constructing an undirected weighted graph from the local energy distribution of the targets in the R-D spectrum, using the echo amplitudes, Doppler indices, and range indices within the energy peaks and their surrounding neighborhoods as node information, and characterizing the scattering characteristics of the targets by setting the weights of the edges between nodes. With the help of GAE, this variable-sized graph structure is further converted into a fixed-dimensional graph feature vector, enabling the effective extraction and vector-form quantization of the electromagnetic scattering information of the targets.

[0015] 3. During the trace-to-track correlation process, in order to quantitatively compare the graph feature vectors of candidate traces and historical traces on the target track, the present invention uses cosine similarity as a similarity metric. Cosine similarity calculation can effectively eliminate the influence of vector magnitude and pay more attention to the direction information of the feature distribution, thereby accurately quantifying the matching degree between graph feature vectors. By calculating the cosine similarity between the graph feature vectors of candidate measurement traces and the graph feature vectors of corresponding frames in the historical track, and combining a multi-frame weighted fusion strategy, the candidate measurement trace with the highest similarity is finally selected as the correct correlation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The specification drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0017] Figure 1 It is an example diagram of the R-D spectrum in the present invention; Figure 2 It is an example diagram of the target foreground in the present invention; Figure 3 It is the structural diagram of the GAE network in the present invention; Figure 4 It is the schematic diagram of the extraction of the graph feature vector in the present invention; Figure 5 It is the schematic diagram of trace-to-track correlation in the present invention; Figure 6 It is the comparison diagram of the track tracking results in the multi-target adjacent scenario in the present invention; Figure 7 It is the comparison diagram of the track tracking results in the sea clutter interference scenario in the present invention; Figure 8 This is the flow chart of the point-track association method of the present invention. Specific embodiments

[0018] It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0019] Embodiment 1 In this embodiment, a point-track association method for ship target based on graph representation learning is provided. As Figure 8 shown in the flow chart of the point-track association method, the specific steps include: S1: Obtain the echo signal containing ship target data and marine environment data, and preprocess the echo signal to obtain the R-D spectrum; S2: Based on the R-D spectrum, divide the target foreground, and use the cells in the target foreground as nodes to construct an undirected weighted graph structure including a node feature matrix and an adjacency matrix; S3: Use the encoder to convert the graph structure into a low-dimensional latent embedding, and then use the decoder to reconstruct the adjacency matrix according to the latent embedding; train the encoder and the decoder. After the training is completed, use the encoder to extract features from the graph structure to obtain the graph feature vector; S4: Obtain the graph feature vectors of the candidate measurement points and the graph feature vectors of the current three frames of track history respectively. At the point-track association gate, calculate the similarity between the graph feature vectors of the candidate measurement points and the graph feature vectors of the current three frames of track history respectively, and select the candidate measurement point with the highest similarity as the point associated with the current track.

[0020] In step S1, a compact ground wave radar is used to obtain the echo signal containing ship target data and marine environment data. Specifically: after the electromagnetic wave of the transmitter interacts with the ship target at sea and the surrounding marine environment, the generated echo signal is received by the receiver.

[0021] The specific method for preprocessing the echo signal is: first convert the echo signal into a digital signal through A / D conversion, and then, through a series of digital signal processing techniques, including pulse compression, beamforming, and Doppler processing, convert the echo signal into an R-D spectrum. As Figure 1 shown, this R-D spectrum is the basic data for subsequent processing, and each resolution cell of the R-D spectrum is represented by a triple composed of distance, Doppler velocity, and echo amplitude.

[0022] In step S2, the specific method for dividing the target foreground is: based on the above R-D spectrum obtained in step S1. As Figure 1As shown, a 5×5 window is selected with the cell where the target peak is located as the center. This window can cover the spatial expansion range of the target echo in the R-D spectrum. The Otsu method is used to determine the optimal amplitude segmentation threshold S between the foreground and background (clutter / noise) within the window. The Otsu method determines the threshold by maximizing the between-class variance between the foreground and background, maximizing the difference between the two categories of foreground and background. Based on the optimal amplitude segmentation threshold S a target foreground segmentation mask is generated. Cells with an amplitude greater than or equal to the optimal amplitude segmentation threshold S are divided into the extended area of the target, that is, the target foreground, as Figure 2 shown.

[0023] In this embodiment, the target foreground segmentation mask is defined as follows: ; where is the echo amplitude of the th cell within the 5×5 window. Cells with are marked as target foreground cells.

[0024] In step S2, for each detected target, the cells in the target foreground are used as nodes in the constructed graph structure. Each node consists of a three-dimensional feature vector composed of Doppler velocity, distance, and echo amplitude, denoted as where is the total number of nodes, is the Doppler velocity index, is the distance index, and is the echo amplitude.

[0025] The method for constructing an undirected weighted graph structure is as follows: First, the node corresponding to the target peak cell in the target foreground is set as the central node and numbered "1", and other nodes are numbered in the order of their relative positions to the central node to retain the spatial layout information of the target scattering region; then the central node is connected to other nodes to construct the graph structure, where the weight of the edge is determined by the amplitude difference between the central node and the connected node, that is: . Finally, an undirected weighted graph structure is constructed, where is the node set, is the edge set, the node feature matrix integrates the features of the graph structure , and the adjacency matrix represents the connection relationship and weight of the edges.

[0026] In step S3, this embodiment uses a Graph Autoencoder (GAE) to extract features from the constructed graph structure. GAE consists of an encoder and a decoder and is an unsupervised learning framework. In the training phase, the encoder processes the input graph structure using a Graph Convolution Network (GCN). The graph convolution network learns the feature representation of nodes by aggregating information from adjacent nodes. Its principle is to perform a convolution operation on the topological structure of the graph so that nodes can obtain the feature information of neighboring nodes.

[0027] Specifically: as Figure 3 shown, in the training phase, the graph structure is input into the encoder. The encoder uses the graph convolution network to (including the node feature matrix and the adjacency matrix ) and converts it into a low-dimensional latent embedding. Its encoding process can be expressed as: ; where is the low-dimensional latent embedding, is the graph convolution network, is the node feature matrix, is the adjacency matrix.

[0028] The decoder reconstructs the adjacency matrix according to the latent embedding , and the reconstruction formula is: ; where is the reconstructed adjacency matrix, is the activation function.

[0029] During the training process, the model is optimized by minimizing the reconstruction loss between the original adjacency matrix and the reconstructed adjacency matrix . The Adam optimization algorithm is used to adjust the model parameters so that the encoder gradually learns effective embeddings to capture the structural features of the graph.

[0030] After training is completed, the encoder is used to extract features from the graph structure. As Figure 4As shown, the encoder adopts a two-layer GCN architecture. First, the first layer of GCN converts the input graph structure into a 64-dimensional hidden embedding through message passing operations, and then uses the Leaky ReLU activation function to introduce non-linearity to enhance the model's ability to capture complex patterns. The second layer of GCN further optimizes the hidden embedding. Based on the hidden embedding generated by the first layer, the second layer of GCN executes the message passing mechanism again, further explores the complex correlation relationships between features by aggregating the features of the node itself and its neighbor nodes, and generates more discriminative and representative node embeddings , where each row of data corresponds to the feature vector of a node in the graph structure. Finally, through global average pooling operation, the feature vectors of all nodes are aggregated to obtain a fixed-length 64-dimensional global graph feature vector , and the calculation formula is: , and this graph feature vector effectively encodes the unique electromagnetic scattering characteristics of the target

[0031] In step S4 Figure 5 is a schematic diagram of point-track association. Inside the point-track association gate, the cosine similarity is used to calculate the similarity between the graph feature vector of the candidate measurement point track and the graph feature vector of the current track history and , their cosine similarity can be calculated by the following formula: ; In this embodiment, the graph feature vector of the candidate measurement point track and the graph feature vectors of the three frames of the current track history are obtained respectively, the cosine similarities between the graph feature vector of the candidate measurement point track and the graph feature vectors of the three frames of the current track history are calculated respectively, and then the comprehensive similarity is obtained through weighted fusion. The calculation formula is: ; where , , are weight coefficients, satisfying >[[]]END]] >[[]]END]] and + + = 1, to highlight the role of the latest target state in the association and at the same time take into account historical information to ensure stability is the graph feature vector of the candidate measurement point track , , are the graph feature vectors of the three frames of the current track history

[0032] If the target state of a certain frame is a prediction rather than a measurement value (that is, there is a lack of effective radar echo and the corresponding graph feature vector), then the corresponding weight coefficient Set it to 0. To enhance the similarity difference, the formula is used to transform the comprehensive similarity, where = is the numerical stability factor to prevent the logarithmic term from diverging when is close to 1. Finally, the candidate measurement point track with the highest similarity is selected as the track associated with the current track.

[0033] The following are the experimental results and analysis of this embodiment: To quantitatively evaluate the performance of point track association, the target tracking duration is used as a key indicator in this experiment. This indicator reflects the ability of the radar system to maintain tracking continuity. The longer the tracking duration, the higher the accuracy of the point track association of the system. In addition, by qualitatively analyzing the association effect under the influence of multi-target proximity and sea clutter interference, the association accuracy and track integrity of different methods in complex environments are compared.

[0034] Quantitative analysis: Statistical analysis is carried out on the average tracking duration of 10 representative targets. The method provided in this embodiment is compared with the traditional Nearest Neighbor Data Association (NNDA) and the advanced Joint Probability Data Association (JPDA) methods. The results are shown in Table 1: Table 1 Comparison of average tracking durations (Unit: minutes, "+" means the tracking durations of multiple track segments are added together)

[0035] The statistical results show that the average tracking duration of the method proposed in this embodiment reaches 135.6 minutes, which is significantly improved compared with NNDA (average 99.7 minutes) and JPDA (average 102.2 minutes), with an average extension of about 32% - 36%, indicating that this method has significant advantages in target tracking continuity.

[0036] Qualitative analysis: 1. Multi-target proximity scenario As Figure 6 shown, in the multi-target proximity case, the traditional NNDA method has a point track association error caused by adjacent target interference at the 70th minute (marked by the blue dot), resulting in the premature termination of the target track; while in this embodiment, the measurement point tracks of adjacent targets are effectively distinguished by using the graph feature vector, achieving correct point track association and increasing the target tracking duration by 88 minutes.

[0037] 2. Sea clutter interference scenario As shown Figure 7 In the sea clutter interference case shown above, due to the influence of sea clutter interference, it is difficult for traditional methods to achieve accurate point-track association in the initial stage, resulting in the track being unable to start in a timely and accurate manner. However, this embodiment can establish the target track in a timely manner and maintain a high association accuracy within the sea clutter interference area, ultimately extending the target tracking duration by 14 minutes (the track segment marked by the blue dashed ellipse).

[0038] The above qualitative analysis cases illustrate that this method has high robustness in complex environments, can achieve accurate point-track association, and ensure continuous and stable tracking of the target track.

[0039] Embodiment 2 In this embodiment, a vessel target point-track association system based on graph representation learning is provided, including the following modules: A data acquisition and preprocessing module, configured to: acquire an echo signal containing vessel target data and marine environment data, and preprocess the echo signal to obtain an R-D spectrum; A graph structure construction module, configured to: based on the R-D spectrum, divide the target foreground, and use the cells in the target foreground as nodes to construct an undirected weighted graph structure including a node feature matrix and an adjacency matrix; A feature extraction module, configured to: use an encoder to convert the graph structure into a low-dimensional latent embedding, and then use a decoder to reconstruct the adjacency matrix according to the latent embedding; train the encoder and the decoder, and after the training is completed, use the encoder to extract features from the graph structure to obtain a graph feature vector; A point-track association module, configured to: respectively obtain the graph feature vectors of candidate measurement points and the graph feature vectors of the last three frames of the current track history, and at the point-track association gate, calculate the similarity between the graph feature vectors of the candidate measurement points and the graph feature vectors of the last three frames of the current track history, and select the candidate measurement point with the highest similarity as the point associated with the current track.

[0040] For those skilled in the art, various changes and modifications can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A ship target point track-track association method based on graph representation learning, characterized in that: include: Acquire the echo signal containing the ship target data and the marine environment data, and pre-process the echo signal to obtain the RD spectrum; Based on the RD spectrum, the target foreground is divided, and the cells in the target foreground are used as nodes to construct an undirected weighted graph structure including a node feature matrix and an adjacency matrix; The graph structure is converted into a low-dimensional latent embedding by using an encoder, and an adjacency matrix is ​​reconstructed according to the latent embedding by using a decoder; the encoder and the decoder are trained, and after the training is completed, the encoder is used to extract features of the graph structure to obtain a graph feature vector; The graph feature vectors of the candidate measurement point tracks and the graph feature vectors of the three historical frames of the current track are obtained respectively. In the point track-track association gate, the similarities of the graph feature vectors of the candidate measurement point tracks and the graph feature vectors of the three historical frames of the current track are calculated respectively, and the candidate measurement point track with the highest similarity is selected as the point track associated with the current track.

2. The method for associating ship target point track and track based on graph representation learning as claimed in claim 1, characterized in that: The preprocessing includes: first converting the echo signal into a digital signal through A / D conversion, and then converting it into an RD spectrum through pulse compression, beam forming and Doppler processing; each resolution unit cell of the RD spectrum is represented by a triple consisting of distance, Doppler velocity and echo amplitude.

3. The method for associating ship target point track and track based on graph representation learning as claimed in claim 1, characterized in that: The method for dividing the target foreground is as follows: in the RD spectrum, a window is selected with the cell where the target peak is located as the center, and the Otsu method is used to determine the optimal amplitude segmentation threshold between the foreground and the background in the window; a target foreground segmentation mask is generated based on the optimal amplitude segmentation threshold, and cells with amplitudes greater than or equal to the optimal amplitude segmentation threshold are divided into target foregrounds.

4. The method for associating ship target point track and track based on graph representation learning as claimed in claim 1, characterized in that: The method for constructing an undirected weighted graph structure is as follows: first, the node corresponding to the target peak cell in the target foreground is set as the central node and numbered as "1", and the other nodes are numbered in order of their relative positions to the central node; then the central node is connected to the other nodes, and finally an undirected weighted graph structure is constructed.

5. The method for associating ship target point track and track based on graph representation learning as claimed in claim 1, characterized in that: During the training phase, the graph structure is input into the encoder, which uses a graph convolutional network to convert the graph structure into a low-dimensional potential embedding. The encoding process can be expressed as: ;in, is a low-dimensional latent embedding, is a graph convolutional network, is the node feature matrix, is the adjacency matrix; The formula for the decoder to reconstruct the adjacency matrix based on the potential embedding is: ;in, is the reconstructed adjacency matrix, is the activation function.

6. The method for associating ship target point track and track based on graph representation learning as claimed in claim 1, characterized in that: The encoder adopts a two-layer GCN architecture; after training, the first layer of GCN converts the input graph structure into a hidden embedding through a message passing operation, and then introduces nonlinearity using a Leaky ReLU activation function; The second layer of GCN executes the message passing mechanism again to generate node embeddings based on the hidden embeddings generated by the first layer; finally, the feature vectors of all nodes are aggregated through the global average pooling operation to obtain the graph feature vector.

7. The method for associating ship target point track and track based on graph representation learning as claimed in claim 1, characterized in that: The similarity between the graph feature vector of the candidate measurement point track and the graph feature vector of the three historical frames of the current track is calculated respectively. The specific method is: the cosine similarity between the graph feature vector of the candidate measurement point track and the graph feature vector of the three historical frames of the current track is calculated respectively, and then the comprehensive similarity is obtained by weighted fusion. The calculation formula is: ;in, , , is the weight coefficient, is the graph feature vector of the candidate measurement point trace, , , It is the graph feature vector of the three historical frames of the current track.

8. A ship target point track-track association system based on graph representation learning, characterized in that: Includes the following modules: The data acquisition and preprocessing module is configured to: acquire an echo signal including ship target data and marine environment data, and preprocess the echo signal to obtain an RD spectrum; The graph structure building module is configured to: divide the target foreground based on the RD spectrum, and use the cells in the target foreground as nodes to build an undirected weighted graph structure including a node feature matrix and an adjacency matrix; The feature extraction module is configured to: use an encoder to convert the graph structure into a low-dimensional potential embedding, and then use a decoder to reconstruct an adjacency matrix according to the potential embedding; train the encoder and the decoder, and after the training is completed, use the encoder to extract features from the graph structure to obtain a graph feature vector; The point track-track association module is configured to: obtain the graph feature vector of the candidate measurement point track and the graph feature vector of the three historical frames of the current track respectively, calculate the similarity of the graph feature vector of the candidate measurement point track and the graph feature vector of the three historical frames of the current track respectively at the point track-track association gate, and select the candidate measurement point track with the highest similarity as the point track associated with the current track.

9. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps in the ship target point track-track association method based on graph representation learning as described in any one of claims 1 to 7 are implemented.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the ship target point track-track association method based on graph representation learning as described in any one of claims 1 to 7 are implemented.

Images